Research And Implementation Of Abnormal Account Detection Technology In Social Network Based On Spark Platform

As an Internet application service emerging in the web 2.0 era,social networking extends offline social activities online,allowing users to socialize online via registered accounts.While social networking platforms promote good online social behavior,the practice of malicious behavior through social network accounts has taken place.Due to the openness and immediacy of social networks,the impact of these malicious acts can spread rapidly and widely.Therefore,the detection of abnormal accounts devoted to malicious behaviors plays an important role in maintaining the security of the network environment.Traditional anomaly detection technology intercepts malicious behavior by maintaining a set of rules.On the one hand,rigid rules can easily be bypassed;on the other hand,the threshold and cost of constructing and maintaining rule sets have been increasing.A new generation web anomaly detection technology based on machine learning technology is expected to bring new developments and breakthroughs for social network platform against attacks.This paper thoroughly analyzes the status quo of anomaly detection technology and the application status of machine learning technology,and proposes an improved automatic encoder model to detect the anomaly accounts of social networks.In addition,according to the large amount of data in social network logs and the requirements of computing performance,this dissertation studies the distributed computing technology in depth.In this paper,we choose Spark distributed computing platform to process the data,which can break the limitation of stand-alone performance,complete the calculation under the mass data more efficiently and accurately,and then finish the training of the more complex and efficient algorithm model.This paper also analyzes the system model and computing platform for the social network anomaly detection,studies the performance tuning of distributed computing,and puts forward the optimization solution to the actual problems,and realizes the reference value to the real production environment.